interep

interep implements penalized generalized estimating equations and structured-sparsity methods to perform interaction analysis of high-dimensional main and interaction effects in repeated measurement data for gene-environment (G × E) interaction studies.

Key Features:

• R package: Implements the described methods as an R package for statistical analysis.
• High-dimensional interaction analysis: Identifies main and interaction effects in settings with many predictors.
• Repeated measurement data handling: Uses a GEE framework to accommodate correlated outcomes from repeated measurements.
• Structured sparsity imposition: Applies structured sparsity assumptions to select relevant main and interaction effects.
• Penalized Generalized Estimating Equations (GEE): Implements penalized GEE methods that combine individual- and group-level penalties for variable selection.
• Alternative individual-level selection methods: Provides alternative approaches that select effects at the individual level without enforcing group-level interaction structures.

Scientific Applications:

• Gene-environment (G × E) interaction studies: Enables detection and estimation of G × E interactions in genetic epidemiology using repeated-measure study designs.
• High-dimensional repeated-measure analysis in bioinformatics: Supports analysis of high-dimensional longitudinal or repeated-measure datasets to identify main and interaction effects relevant to disease mechanisms.

Methodology:

The methodology is based on the penalized GEE approach proposed by Zhou et al. (2019), employing a mixture of individual- and group-level penalties to enforce structured sparsity for selection of main and interaction effects in high-dimensional repeated-measure datasets.